ChatGPT can now see, hear, and speak

OpenAI is beginning to roll out voice and image capabilities in ChatGPT, letting users hold spoken conversations with the assistant and show it pictures to discuss. The features are rolling out to Plus and Enterprise users over two weeks, with voice on iOS and Android as an opt-in setting and images on all platforms. The voice feature uses a new text-to-speech model that can generate human-like audio from text and a few seconds of sample speech, and it offers five voices created with professional voice actors. OpenAI says it is deploying the features gradually to refine risk mitigations and limit misuse, such as impersonation of public figures.